pan-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Pan-users] Body pane is wrongly resizing pan's top-level window


From: Duncan
Subject: Re: [Pan-users] Body pane is wrongly resizing pan's top-level window
Date: Sat, 4 Jul 2015 04:42:05 +0000 (UTC)
User-agent: Pan/0.140 (Chocolate Salty Balls; GIT af87825)

walt posted on Fri, 03 Jul 2015 15:54:43 -0700 as excerpted:

> On Fri, 3 Jul 2015 06:22:05 +0000 (UTC)
> Duncan <address@hidden> wrote:
> 
> <magnum opus snipped for brevity while I take a few days to absorb it>
> ;-)
> 
> I wanted to post a sample 'problem' jpeg file that you could view in pan
> to see if it causes the same problem for you,

This suggests that at least the problem jpeg itself wasn't too 
"sensitive" to post.  There does, however, remain the issue of whether 
the group it was originally in might be "sensitive", as well as whether 
it was a public access group at all, of course.

But given that I do have reasonably good binary-group access, now, even 
if I don't actually use it much, if the original post is on a publicly 
distributed group or perhaps on a gmane.* list-group, assuming the group 
itself isn't too sensitive, consider posting the original problem post 
information: group, subject/title, author, date/time posted (UTC or 
convertable), and message-id.  Chances are reasonable that I have access 
to the post too, and can thus answer your question of whether it triggers 
issues for me here, as well.

Of course you may not wish to publish it to the list.  If you wish, mail 
me directly.  (If you're following the list on gmane my posting address 
will of course have gmane original email obfuscation, but the gmane-
obfuscated address is actually valid as well, it simply forwards thru 
gmane, assuming of course it doesn't get caught by gmane despamming.)

Tho of course it's understandable if you consider the original group too 
sensitive, as well, even for direct emailing.  Far be it from me to 
pressure someone with privacy concerns into overriding them! =:^)  But 
that you were already working on posting the image does suggest that the 
image itself, at least, isn't overly sensitive, a reasonable hint that 
the group it was on and the subject of the post aren't too sensitive, as 
well.

> but I just discovered that I can't do it.
> 
> I used pan to save a 'problem' jpeg to disk, and then I used pan to post
> the same (unedited) jpeg to a test group.
> 
> I was amazed to find that pan behaved normally when viewing the problem
> jpeg that I posted myself:  the body pane added a scroll bar so the
> entire jpeg could be seen without expanding the top-level window, just
> as expected.
> 
> I went to pan's article-cache and used uudeview to extract both the
> original article and the one I posted.  The two jpeg files are bitwise
> identical according to cmp.
> 
> This is the most desirable outcome from any experiment:  a surprise :)

Indeed! =:^)

Meanwhile, you're probably well beyond this already, but sometimes it's 
nice to get independent confirmation of one's troubleshooting techniques 
and operational hypothesis, so...

That the files are bitwise identical does indicate that it's not the 
image that's the problem, but rather, something about the way they were 
posted.

Given a basic understanding of news/mail RFCs and message format basics...

One hypothesis would be that a particular combination of posting format 
and attachment handling, perhaps MIME multipart/mixed with text/html for 
the text part, and a specific content-disposition for the attachment, 
triggers the bug.

Another would be that a particular combination of uncommon characters in 
the text content triggers the bug.  I've seen other bugs triggered due to 
uncommon characters -- one was that pan simply wouldn't display the rest 
of the... was it only the paragraph, or the entire message... after a 
particular character, at least in particular charsets (UTF-8 I believe 
that was).  IDR whether that one was fixed or not.

Meanwhile, I /do/ have saved a particular bug-triggering post, with the 
bug actually being a pan crash with segfault, due to an empty MIME part.  
(In that particular case, the message is on gmane, with, based on the 
gmane headers, the empty MIME part due to gmane scrubbing malware from 
the message but leaving the enclosing message-part intact, but empty.  
Pan apparently can't deal with an entirely empty MIME message part.)  
That bug has NOT been fixed.  IIRC I posted it to the dev list but never 
got around to filing a bug for it... I should.)

> I'll keep chewing on this surprise while I continue to decode Duncan's
> post...

LOL.  As I warned, it did take the scenic route.  Some of it's certainly 
nerd boasting, but I know from personal experience that exactly that sort 
of side-content has often lead to the most unexpected and powerful 
Eureka! moments, here, changing the way I approach something or giving me 
entirely new insights into things I'm dealing with that may not be 
related to the nominal subject at hand, at all.  If I can return the 
favor, passing on a bit of unexpected side-channel help to others as I've 
received it, and if it's even half as useful to them as the bits I 
gleaned were useful to me, great!  In fact, it''s exactly that sort of 
thing that has kept me coming back to the lists and groups, year after 
year, now decade after decade.

But, as I said, it did get back around to topical, since it's quite 
possible that my window rules for pan do constrain the effect of such 
bugs to some extent, tho as explained I don't believe they'd entirely 
stop the effect, unless there's some unexpected third factor involved 
with the bug as well.

Which is why, if possible, I'd like a reference to the for you bug-
triggering original post, so I can see it in its native context, and see 
exactly what sort of behavior it triggers... or not... here.

-- 
Duncan - List replies preferred.   No HTML msgs.
"Every nonfree program has a lord, a master --
and if you use the program, he is your master."  Richard Stallman




reply via email to

[Prev in Thread] Current Thread [Next in Thread]